...
At the beginning, Odysseus translates the query in a logical graph. This logical graph is preprocessed and then partitioned. The result of the Partition phase is a collection of query parts, which are individually modified in the Modification phase. However, the Modification-phase is optional an can be omitted. After that, the Allocation phase decides, which query part should be executed on which OdysseusNode. Finally, the query parts are transmitted to the selected nodes. Preprocess, Partition, Modification, Allocation and Postprocess provide multiple strategies the user can select from (individually for each query, if needed). For this, the Additionally, some strategies can be parameterised . The user has to use specific OdysseusScript keywords. One example for query distribution is as follows:
| Code Block |
|---|
| language | java |
|---|
| linenumbers | true |
|---|
|
/// Indication that we want to distribute queries
#CONFIG DISTRIBUTE true
#NODE_PREPROCESS SOURCE /// Select Source strategy for pre-processing
#NODE_PARTITION OPERATORCLOUD /// Select OperatorCloud strategy for Partition phase
#NODE_MODIFICATION REPLICATION 2 /// Select Replication strategy for Modification phase (with replication degree 2)
#NODE_ALLOCATION USER /// Select User-strategy for Allocation phase
#NODE_POSTPROCESSOR LOCALSOURCES /// Select multiple strategies for post-processing
#NODE_POSTPROCESSOR LOCALSINKS
#NODE_POSTPROCESSOR MERGE
/// "Typical" query definition here
#METADATA TimeInterval
#PARSER PQL
#ADDQUERY
<Query> |
...
Modification
| Strategy name | Description |
|---|
Replicationreplication <degree> | Query parts are replicated (with replicationdegree degree), executed multiple times. See Replication. | Fragmentation | Query parts are replicated, but each copy retrieves a part of the incoming data stream. See Fragmentation. |
Allocation
| Strategy name | Description |
|---|
direct | All query parts are assigned to the specified OdysseusNode. | Code Block |
|---|
#NODE_ALLOCATE DIRECT MyNode1 |
|
querycount | Query parts are assigned to nodes with the least count of queries. |
roundrobin | Query parts are assigned in order. |
user | Query parts are assigned to user-specified nodes. For this, each logical operator has a DESTINATION-Parameter in PQL. Two operators which have the same DESTINATION-value are assigned to the same node. |
Postprocess
| Strategy name | Description |
|---|
localsink | Logical sink operators are staying local (on the distributing node) overriding allocation. Useful if the user wants to have the last operator to check the data stream results. |
localsource | Logical source operators are staying local (on the distribution node) overriding allocation. Useful if the user do not want to share its source operators to other nodes. |
merge | Two adjacent query parts which are assigned to the same node are merged to omit unneeded network-transmission operators. |